Interaction of liming and long-term fertilization increased crop yield and phosphorus use efficiency (PUE) through mediating exchangeable cations in acidic soil under wheat-maize cropping system

Inorganic fertilizers, particularly urea and other ammoniacal sources, are widely used to achieve high crop yields but can lead to significant soil acidification through nitrification-driven proton release. Soil pH directly and indirectly regulates soil biochemical properties, nutrient availability, and plant growth. In acidic soils, phosphorus (P) availability is reduced due to fixation with Al and Fe oxides, contributing to low P use efficiency (PUE) and limiting crop yields in southern China. Liming is considered an effective practice to ameliorate soil acidity, potentially increasing P availability and crop productivity. However, P solubility and adsorption dynamics with pH can be complex, and field responses vary. The main objectives of this study were to investigate the relationships between soil pH, PUE, and crop yield under long-term liming and fertilization in acidic soil, and to quantitatively assess the factors limiting PUE and crop yield under a wheat–maize rotation.

Prior studies show that nitrogen fertilization induces soil acidification, depleting base cations (Ca2+, Mg2+, K+, Na+) and increasing soluble toxic metals such as Al, Fe, and Mn. Soil P availability is highly sensitive to pH; in acidic soils, P is often fixed by Fe and Al oxides, reducing plant uptake. Liming has been reported to mitigate acidification, raise pH, increase base saturation, and can increase plant-available P, though excessive liming may lead to P fixation with Ca and reduced soil available P. Organic amendments (manure, straw) can also raise pH and enhance P cycling enzymes, improving PUE. Theoretical and experimental work on P adsorption indicates maximum adsorption at low pH (around 4) for Al/Fe oxides, supporting the rationale for liming to reduce Al/Fe activity and improve P uptake. Regional evidence indicates intensified acidification in Chinese croplands due to long-term N inputs and atmospheric N and S deposition, underscoring the need for mitigation strategies.

Experimental site: Long-term field trial initiated in 1990 at the National Observation and Research Station of Farmland Ecosystem, Qiyang County, Hunan, China (26°45′42″ N, 111°52′32″ E). Subtropical monsoon climate, mean annual temperature 17.8 °C, rainfall 1290 mm. Soil: Eutric Cambisol (WRB), Inceptisol (USDA), light loam, red soil (China). Texture: 43.86% clay, 31.86% silt, 24.28% sand. Initial (1990, 0–20 cm): pH 5.7, SOC 7.9 g kg−1, TN 1.07 g kg−1, AN 79 mg kg−1, TP 0.45 g kg−1, AP 14.0 mg kg−1, TK 13.7 g kg−1, AK 104 mg kg−1. Experimental design: Winter wheat–summer maize rotation; split-plot design with two true replicates; a third pseudo-replication sampled within one replicate. Plot size 20 m × 5 m with 20-cm cement baffles. Seven treatments: (1) CK (no fertilization), (2) NP (inorganic N and P), (3) NPK (inorganic N, P, K), (4) NPKS (NPK + 50% straw return of wheat and maize), (5) NPCa (NP + lime), (6) NPKCa (NPK + lime), (7) NPKSCa (NPKS + lime). Annual fertilizer rates: urea 150 kg N ha−1, calcium superphosphate 120 kg P2O5 ha−1, potassium chloride 120 kg K2O ha−1. Fertilizers applied before sowing; 30% to wheat, 70% to maize. Lime (quicklime, CaO) applied at 2550 kg ha−1 in Oct 2010 and 1500 kg ha−1 in Oct 2014 in limed treatments. In NPKS/NPKSCa, 50% of aboveground straw incorporated (excess nutrient inputs via straw not balanced). Cropping: Wheat (Xiangmai), 63 kg ha−1 (~160 seeds m−2); Maize (Yedan-13), 60,000 seeds ha−1. No irrigation. Pest control: Omethoate and Carbofuran for wheat aphid; control of maize borers; Glyphosate for post-harvest weed control. Grains and straw removed; stubble (~6 cm) and roots left in soil. Sampling and analyses: Annual topsoil (0–20 cm) samples collected post-maize harvest 2012–2018; five cores per plot composited, air-dried, sieved (0.25 mm). Soil pH measured in 2.5:1 water:soil suspension (glass electrode). SOC by dichromate oxidation; TN (Black, 1965); TP (Murphy & Riley, 1964); TK (Knudsen et al., 1982). Available N, P, K following Lu (2000), Olsen (1954), and Page et al. (1982). Exchangeable Ca2+ and Mg2+ extracted with 1 M NH4OAc (pH 7) and measured by AAS; exchangeable Al3+ by NaOH neutralization titration after 0.1 M BaCl2 extraction. Crop samples (grain and straw) dried (105 °C for 0.5 h, then 70 °C to constant weight), ground, digested (H2SO4–H2O2 at 270 °C); P determined by vanadomolybdate yellow method. Calculations: Phosphorus use efficiency (PUE; kg kg−1) defined as P agronomic efficiency: PUE = (YF − Y0)/F, where YF is annual above-ground biomass yield under fertilization, Y0 is yield under CK, and F is annual P input (kg ha−1). Statistics: One-way ANOVA for treatment effects; two-way ANOVA for treatment × year interactions; Tukey’s HSD at P = 0.05 (Statistix 8.1). Linear regressions for relationships among soil properties, PUE, and yield. Boosted Regression Trees (BRT) using R (gbm, v3.3.3), tenfold cross-validation, tree complexity 5, learning rate 0.005, modeling relative influence of predictors on annual crop yield.

- Soil pH: Long-term inorganic fertilization decreased soil pH over time; liming with fertilization increased pH, though CK had the highest average pH (5.77). Relative to CK, mean pH decreases were 25.4% (NP), 26.0% (NPK), 24.2% (NPKS), 14.8% (NPCa), 12.1% (NPKCa), and 14.7% (NPKSCa). - Soil C and N: SOC, TN, and AN were higher in all fertilized treatments than CK. Average increases vs CK: SOC by 24.0% (NP), 38.8% (NPK), 35.7% (NPKS), 33.2% (NPCa), 39.3% (NPKCa), 29.8% (NPKSCa). TN by 18.3%, 25.7%, 26.0%, 20.3%, 20.8%, and 23.6%, respectively. AN by 38.1%, 49.7%, 32.0%, 40.7%, 25.2%, and 32.3%, respectively. - Soil P: Total P increased vs CK by 107% (NP), 130% (NPK), 128% (NPKS), 118% (NPCa), 113% (NPKCa), 95% (NPKSCa). Available P increased vs CK by 1668%, 1709%, 1954%, 1699%, 1315%, and 1325% for NP, NPK, NPKS, NPCa, NPKCa, NPKSCa, respectively. Available P tended to be higher under NPK/NPKS than under limed NPKCa/NPKSCa. - Exchangeable cations: Liming increased exchangeable Ca2+ and Mg2+ and decreased Al3+ vs unlimed fertilizer treatments. Average exchangeable Ca2+ (6.8 cmol kg−1) was highest in CK. Relative to CK, exchangeable Ca2+ decreased by 39% (NP), 37% (NPK), 48% (NPKS), 11% (NPCa), 10% (NPKCa), 15% (NPKSCa). Exchangeable Mg2+ increased by 16.4% under NP, but decreased by 38% (NPK), 53% (NPKS), 18.7% (NPCa), 42.3% (NPKCa), 21.2% (NPKSCa). Exchangeable Al3+ increased vs CK by 1576% (NP), 1518% (NPK), 1308% (NPKS), 499% (NPCa), 430% (NPKCa), 491% (NPKSCa). - Crop yields: Fertilization with liming significantly increased wheat and maize yields compared to unlimed fertilization. Mean increases vs CK: wheat +138% (NP), +213% (NPK), +198% (NPKS), +547% (NPCa), +688% (NPKCa), +626% (NPKSCa); maize +687%, +1887%, +1651%, +2605%, +5047%, +5077%, respectively. Highest yields under NPKCa and NPKSCa. - P uptake and PUE: Mean P uptake increases vs CK were +154% (NP), +461% (NPK), +472% (NPKS), +717% (NPCa), +1168% (NPKCa), +1236% (NPKSCa). Mean PUE (2012–2018): 20.7 (NP), 66.2 (NPK), 64.4 (NPKS), 105.1 (NPCa), 187.6 (NPKCa), 185.0 kg kg−1 (NPKSCa). Relative to NP, PUE increased by 220% (NPK), 212% (NPKS), 409% (NPCa), 807% (NPKCa), 795% (NPKSCa). - Relationships (linear regressions): pH vs exchangeable Ca2+: R2 = 0.666, p < 0.001 (positive). pH vs exchangeable Mg2+: R2 = 0.019, p > 0.05 (ns). pH vs TN: R2 = 0.218, p < 0.001 (negative). pH vs exchangeable Al3+: R2 = 0.817, p < 0.001 (negative). PUE vs pH: R2 = 0.396, p < 0.001 (positive). PUE vs exchangeable Ca2+: R2 = 0.435, p < 0.001 (positive). PUE vs exchangeable Al3+: R2 = 0.488, p < 0.001 (negative). Annual crop yield vs pH: R2 = 0.399, p < 0.001 (positive). Yield vs PUE: R2 = 0.956, p < 0.001 (positive). - BRT analysis: Most influential predictors of annual crop yield were exchangeable Ca2+ (33.5%), soil pH (23.9%), exchangeable Al3+ (11.6%), available N (7.7%), available P (6.6%); Mg2+, TN, TP, SOC each <5%. Model fit: R2 = 0.988 between observed and predicted yields.

The study demonstrates that long-term inorganic fertilization without liming exacerbates soil acidification, depletes base cations, and elevates exchangeable Al3+, which collectively suppress P availability, P use efficiency, and crop yields in acidic red soils. Liming effectively counteracted acidification by raising pH, increasing exchangeable Ca2+ and Mg2+, and reducing Al3+. These changes improved P uptake and P agronomic efficiency, translating into substantial yield gains for both wheat and maize, particularly when liming was combined with balanced NPK fertilization and straw return (NPKCa, NPKSCa). Strong positive associations of pH and exchangeable Ca2+ with PUE and yield, and negative associations with Al3+, confirm that alleviating acidity and Al toxicity is central to improving nutrient use efficiency and productivity. While liming sometimes reduced soil test available P compared to unlimed NPK/NPKS—likely due to Ca-associated P fixation—the net effect on plant P uptake and PUE remained positive, suggesting improved root growth, microbial activity, and altered sorption dynamics outweighed any decrease in extractable P. BRT results further highlight exchangeable Ca2+ and pH as primary drivers of yield under these conditions, indicating that mediating the soil cation environment is critical for sustaining productivity in acid croplands.

Long-term fertilization without liming decreased crop yield and PUE due to intensified soil acidification. Quicklime application significantly increased soil pH and base cations (Ca2+, Mg2+) while reducing exchangeable Al3+, thereby enhancing P uptake, PUE, and grain yields. The largest improvements occurred under NPKCa and NPKSCa, where liming and straw incorporation together mitigated acidity and improved soil quality. Although liming reduced soil available P compared to unlimed NPK/NPKS, overall plant P uptake and PUE were still greater with liming. Exchangeable Ca2+, soil pH, exchangeable Al3+, and available N were identified as the most influential factors for annual crop yield. Combining mineral fertilizers with straw return and periodic liming is an effective strategy to achieve high yields and PUE in acidic soils under wheat–maize rotations.

The experimental design included only two true replicates plus one pseudo-replication, which may increase the risk of Type I error and limit statistical robustness. The findings are from a single long-term site with specific soil and climate conditions, which may constrain generalizability to other regions or soil types. Temporal variability indicated some inconsistencies in exchangeable cation trends across years.